@d19fe8

HCI Prof @cmuhcii | ...now at https://xitter.cf/t.co/2DGRG3mZmC

 Pittsburgh, PA
thecoalalab.com
Joined January 2014
  • Tweets1.4K
  • Following779
  • Followers2.1K
36Media
Moving to bsky! (@kenholstein.bsky.social) I won't be checking Twitter often, but will keep this account open.
3
762
Microsoft Research NYC is hiring a researcher in the space of AI and society!
13
41
2
305
56,020
I'm seeking a postdoc to work with me and @d19fe8 on evaluating AI/ML decision support for human experts: statmodeling.stat.columbia.e… P.S. I'll be at NeurIPS Thurs-Mon. Happy to talk about this position or related mutual interests! Please retweet! 🙏
2
21
1
56
8,951
Ken Holstein retweeted
Maia Jacobs@maia_jacobs12 Nov 2024
I am so proud of this research collaboration with @conniewchau and @thenetworkaadv, looking at how we can support frontline workers within gender-based violence organizations. This work is more important than ever, and I'm humbled to be a part of it. CSCW talk is tomorrow @ 9am.
Connie W. Chau@conniewchau12 Nov 2024
‼️#cscw2024 TOMORROW Nov. 13th @ 9am in Session 1c: Care for the Caregivers in Central 3‼️ I'll be presenting the highlights of the study that cemented an ongoing multiyear partnership with @thenetworkaadv (BIG thanks to Colleen Norton & Denise Huang, my community co-authors) 🥰
Title slide for a paper Connie is presenting at CSCW 2024 -- "Envisioning the Future of Burnout Support: Understanding Frontline Workers' Experiences in Nonprofit Gender-Based Violence Organizations".

ALT Title slide for a paper Connie is presenting at CSCW 2024 -- "Envisioning the Future of Burnout Support: Understanding Frontline Workers' Experiences in Nonprofit Gender-Based Violence Organizations".

3
3
26
2,828
If you’re at #CSCW2024, join our SIG on "Collaboratively Designing and Evaluating Responsible AI Interventions"! We’ll explore how CSCW researchers can contribute to RAI research and practice. No need to register—just drop by from 2:30 to 4:00 pm CST in Room Cartago!
Collaboratively designing and evaluating responsible AI interventions, a Special Interest Group at CSCW 2024.

ALT Collaboratively designing and evaluating responsible AI interventions, a Special Interest Group at CSCW 2024.

1
8
49
3,966
Ken Holstein retweeted
Hal Daumé III@haldaume35 Nov 2024
The AI Interdisciplinary Institute at @UofMaryland (AIM) is hiring 40 new faculty members in all areas of AI, particularly: - accessibility, - sustainability, - social justice, and - learning; building on computational, humanistic, or social scientific approaches to AI. >
3
46
1
175
23,463
Have you ever worried that no one would respond to your message? Our #CSCW2024 paper proposes a commitment-based design in online groups to address this: arxiv.org/abs/2410.23267 (with @zhangyt0704 and @msbernst )
Three texts from the same person. The first reads "Do you have advice on what I should put I the thread promoting commit on twitter?" The second reads "I could use some help." After those texts is the read receipt. A third, unread, message follows "Guys????"

ALT Three texts from the same person. The first reads "Do you have advice on what I should put I the thread promoting commit on twitter?" The second reads "I could use some help." After those texts is the read receipt. A third, unread, message follows "Guys????"

8
22
4
109
30,772
Our new preprint shows that using chain of thought prompting can result in a significant drop in the performance of AI models on tasks where getting people to express their thoughts reduces human performance
Ryan Liu @ COLM 26
@theryanliu30 Oct 2024
Is encouraging LLMs to reason through a task always beneficial?🤔 NO🛑- inspired by when verbal thinking makes humans worse at tasks, we predict when CoT impairs LLMs & find 3 types of failure cases. In one OpenAI o1 preview accuracy drops 36.3% compared to GPT-4o zero-shot!😱
5
32
3,841
Ken Holstein retweeted
Myra Cheng@chengmyra121 Oct 2024
A snippet of the work we did over the summer during my internship with MSR FATE Montreal—we highlight the need for more work investigating the impacts of anthropomorphic AI systems, which is critical to understanding the impact of genAI systems: arxiv.org/pdf/2410.08526
3
20
2
118
28,365
If you're at #AIES join oral session 1 as @anna_kawakami presents our (@achould) work on multi-actor RAI artifacts in "Do Responsible AI Artifacts Advance Stakeholder Goals? Four Key Barriers Perceived by Legal and Civil Stakeholders" @AIESConf #aies24
3
41
1
45
9,992
Ken Holstein retweeted
Hong Shen@hongshenus20 Oct 2024
Thrilled to share that we got the Best Paper Award this year at HCOMP! This paper looks at how users’ identities and experiences might impact their rating of harmful algorithmic biases. See link here for the full paper: ojs.aaai.org/index.php/HCOMP…
CMU Human-Computer Interaction Institute@cmuhcii18 Oct 2024
Looking forward to it, including "Investigating What Factors Influence Users’ Rating of Harmful Algorithmic Bias & Discrimination" by @sendgoodcheers, Jiayin Zhi, @wes_deng, Jaimie Lee, Sizhe Zhang, @meslami2, @d19fe8, @jas0nh0ng, @tianshi_li & @hongshenus (a Best Paper Nominee)
3
12
83
9,383
I’ll be at AIES this week, to present our work with @struthious and @andrewthesmart on why imagining future forms of intelligence requires more specific language than ‘AGI’. Ping me if you’d like to chat! arxiv.org/pdf/2401.13142
3
14
1,222
HUGE congrats to @domoritz!! It's been such a pleasure and honor to work with you since the early days of @vega_vis. Here's to many more years of conversation and collaboration 🍻
Interactive Data Lab@uwdata15 Oct 2024
Congratulations to IDL alum @domoritz for winning a VGTC Significant New Researcher award!! A premier honor for early career researchers in visualization!
4
53
3,406
I’m excited to announce that I'm recruiting a PhD student in Human-Centered Computing (HCC) at @UMBC! If you're passionate about building AI systems, the future of work, or community-engaged research, apply to our HCC PhD program! Why UMBC, you ask? Let me tell you! (1/5)
4
99
4
256
30,138
Ken Holstein retweeted
Tom McCoy@RTomMcCoy10 Oct 2024
🤖🧠NOW OUT IN PNAS🧠🤖 Language models show many surprising behaviors. E.g., they can count 30 items more easily than 29 In Embers of Autoregression, we explain such effects by analyzing what LMs are trained to do pnas.org/doi/10.1073/pnas.23… Major updates since the preprint! 1/n
At the top is the title of the paper: "Embers of autoregression show how large language models are shaped by the problem they are trained to solve". Below on the left is a screenshot of ChatGPT being asked to count how many words are in a list. The correct answer is 29, but it says 30. Next to it is a plot showing ChatGPT's accuracy at counting elements in a list; in general, it does well on multiples of 10 but poorly on other numbers. The explanation offered at the bottom of the image is: In training sets, round numbers are much more common than other numbers.

ALT At the top is the title of the paper: "Embers of autoregression show how large language models are shaped by the problem they are trained to solve". Below on the left is a screenshot of ChatGPT being asked to count how many words are in a list. The correct answer is 29, but it says 30. Next to it is a plot showing ChatGPT's accuracy at counting elements in a list; in general, it does well on multiples of 10 but poorly on other numbers. The explanation offered at the bottom of the image is: In training sets, round numbers are much more common than other numbers.

9
82
11
359
56,361
How accurate & ethical are claims of #ArtificialIntelligence outperforming clinicians? "Constructive ethical guidance that can benefit authors, journal editors, & peer reviewers when reporting and evaluating findings in studies comparing AI to physician performance" is offered in this commentary from @JojannekeDrogt, @JongsmaKr et al. from @UMCUtrecht. nature.com/articles/s41746-0…
22
2
43
6,572
Queenie has been the foundation of the @cmuhcii program for its whole existence -- amazing impact on HCI as a field through development and support of everyone who has graduated from our program!
CMU Human-Computer Interaction Institute@cmuhcii2 Oct 2024
Please join us in congratulating longtime staff member, Queenie Kravitz, on her retirement today. She started @CarnegieMellon in 1993 and the HCII in 2004, and as graduate program coordinator certified our very first HCI PhD and master's degrees. Congrats, Queenie! #CMUhcii
2
1
57
4,593
Can recombining aspects of analogous papers and your own interests help you find your next research project? Introducing Scideator 🧪💡, a mixed-initiative tool for scientific ideation that works with you to generate ideas 🔀 and evaluate/improve their novelty 🔍.
5
26
7
88
18,258
📢If you're applying to @SCSatCMU PhD and come from a traditionally underrepresented group, current students are here to help! Apply to the Graduate Applicant Support Program (GASP) by Oct. 8, 2024 to receive feedback on your application materials: cs.cmu.edu/~gasp
gasp flyer

ALT gasp flyer

1
47
3
87
10,432
Excited to have joined @BerkeleyDataSci and @UCBStatistics as an Assistant Professor!
Berkeley Computing, Data Science, and Society@BerkeleyCDSS11 Sep 2024
UC Berkeley’s College of Computing, Data Science, and Society (CDSS) appointed two new chairs and welcomed seven new faculty in statistics, electrical engineering and computer sciences, and computational precision health to its community this fall! 🔗cdss.berkeley.edu/news/cdss-…
1
29
1,246
Load more